Sentiment, reach, and quality: A machine learning approach to social media analytics in food and beverage industry

This paper investigates the application of advanced machine learning techniques to assess food and beverage influencers across major social media platforms within the Indian digital ecosystem. Using data from 50 prominent influencers representing diverse content genres and audience demographics, we developed predictive models to evaluate influencer performance and brand partnership potential. Key features such as comment-to-like ratio, content originality, sentiment polarity, and sharing behavior were analyzed to identify factors influencing consumer engagement and conversion. Experimental results demonstrated that regionally relevant, authentic content outperformed generic promotions in driving user interaction and purchase intent. The proposed linear regression model achieved an accuracy of 92.3% in forecasting engagement patterns, while a random forest-based approach yielded 84.7% accuracy in predicting conversion outcomes. These models exhibited strong generalization on unseen influencer data, validating their practical application in digital marketing strategies. This research offers a data-driven framework to enhance influencer selection, campaign design, and performance measurement for food and beverage brands operating in India.